DHS COMPARATIVE REPORTS 48

Transcription

1 ABSOLUTE POVERTY, FERTILITY PREFERENCES, AND FAMILY PLANNING USE IN FP2020 FOCUS COUNTRIES DHS COMPARATIVE REPORTS 48 AUGUST 2018 This publication was produced for review by the United States Agency for International Development (USAID). The report was prepared by Sarah Staveteig, Tesfayi Gebreselassie, and Kathryn T. Kampa.

2

3 DHS Comparative Reports No. 48 Absolute Poverty, Fertility Preferences, and Family Planning Use in FP2020 Focus Countries Sarah Staveteig 1,2 Tesfayi Gebreselassie 3 Kathryn T. Kampa 3 ICF Rockville, Maryland, USA August Avenir Health 2 The DHS Program 3 ICF Corresponding author: Sarah Staveteig, The DHS Program, 530 Gaither Road, Suite 500, Rockville, MD 20850, USA; phone: ; fax: ; Sarah.Staveteig@icf.com

4 Acknowledgments: We thank Shea Rutstein for a valuable review of our draft manuscript; Tom Pullum, Courtney Allen, and Trevor Croft for assistance with variable harmonization; Anne Morse for assistance compiling regression result tables; and Chris Gramer for assistance with chart legends. Editor: Bryant Robey Document Production: Natalie Shattuck Cover Design: Chris Gramer Map: Tom Fish This study was carried out with support provided by the United States Agency for International Development (USAID) through The DHS Program (#AID-OAA-C ). The views expressed are those of the authors and do not necessarily reflect the views of USAID or the United States Government. The DHS Program assists countries worldwide in the collection and use of data to monitor and evaluate population, health, and nutrition programs. Additional information about The DHS Program can be obtained from ICF, 530 Gaither Road, Suite 500, Rockville, MD USA; telephone: , fax: , info@dhsprogram.com, internet: Recommended citation: Staveteig, Sarah, Tesfayi Gebreselassie, and Kathryn T. Kampa Absolute Poverty, Fertility Preferences, and Family Planning Use in FP2020 Focus Countries. DHS Comparative Reports No. 48. Rockville, Maryland, USA: ICF.

5 CONTENTS TABLES... v FIGURES... vii PREFACE... ix ABSTRACT... xi 1 INTRODUCTION Study Objectives Background Fertility preferences Family planning: Levels and trends Global poverty: Levels and trends Family planning outcomes and poverty DATA AND METHODS Country and Survey Selection Key Outcome Indicators Absolute Poverty Measurement Rationale Previous approaches Our approach Analysis RESULTS Absolute Poverty among Married Women: Levels and Trends Fertility Preferences Mean ideal family size Non-numeric ideal fertility preferences Modern Contraceptive Prevalence Contraceptive Method Mix Levels Trends Decadal changes Demand Satisfied for Modern Methods DISCUSSION AND CONCLUSIONS Absolute Poverty Fertility Preferences Contraceptive Method Mix Modern Contraceptive Prevalence and Demand Satisfied by Modern Methods Overall Conclusions REFERENCES APPENDIX A ABSOLUTE POVERTY COMPONENTS AMONG THE STUDY POPULATION APPENDIX B DISTRIBUTIONS AND ABSOLUTE DECADAL CHANGES IN OUTCOME MEASURES iii

6

7 TABLES Table 2.1 Surveys included in the analysis Table 3.1 Average relative decadal changes in poverty composition, by country Table 3.2 Average relative decadal changes in ideal number of children, by absolute poverty level and country Table 3.3 Average relative decadal changes in non-numeric fertility preferences, by absolute poverty level and country Table 3.4 Average relative decadal changes in modern contraceptive prevalence, by absolute poverty level and country Table 3.5a Average relative decadal changes in long-term modern methods as a percentage of family planning use, by absolute poverty level and country Table 3.5b Average relative decadal changes in short-term modern methods as a percentage of family planning use, by absolute poverty level and country Table 3.5c Average relative decadal changes in traditional methods as a percentage of family planning use, by absolute poverty level and country Table 3.5d Average relative decadal changes in folkloric methods as a percentage of all family planning use, by absolute poverty level and country Table 3.6 Average relative decadal changes in demand satisfied for modern methods, by absolute poverty level and country Table A.1 Levels of four deprivations and of asset poverty, married women Table B.1 Absolute poverty levels among married women, by survey Table B.2 Average absolute decadal changes in poverty composition, by country Table B.3 Mean ideal number of children by absolute poverty level and survey Table B.4 Average absolute decadal changes in ideal number of children, by absolute poverty level and country Table B.5 Percentage of respondents reporting a non-numeric ideal number of children, by absolute poverty level and survey Table B.6 Average absolute decadal changes in non-numeric fertility preferences, by absolute poverty level and country Table B.7 Modern contraceptive prevalence by absolute poverty level and survey Table B.8 Average absolute decadal changes in modern contraceptive prevalence, by absolute poverty level and country Table B.9 Method mix by absolute poverty level and survey Table B.10 Average absolute decadal changes in long-term modern methods as a percentage of family planning use, by absolute poverty level and country Table B.11 Average absolute decadal changes in short-term modern methods as a percentage of family planning use, by absolute poverty level and country Table B.12 Average absolute decadal changes in traditional methods as a percentage of family planning use, by absolute poverty level and country Table B.13 Average absolute decadal changes in folkloric methods as a percentage of family planning use, by absolute poverty level and country Table B.14 Demand satisfied for modern methods by absolute poverty level and survey Table B.15 Average absolute decadal changes in demand satisfied for modern methods, by absolute poverty level and country v

8

9 FIGURES Figure 2.1 Study countries... 9 Figure 2.2 Definitions of absolute poverty groups used in the study Figure 3.1a Absolute poverty composition of women currently in union, Central and Western Africa Figure 3.1b Absolute poverty composition of women currently in union, Eastern and Southern Africa Figure 3.1c Absolute poverty composition of women currently in union, South Asia, Southeast Asia, and Other Areas Figure 3.2a Mean ideal number of children by absolute poverty level and survey year, Central and Western Africa Figure 3.2b Mean ideal number of children by absolute poverty level, Eastern and Southern Africa Figure 3.2c Mean ideal number of children by absolute poverty level, South Asia, Southeast Asia, and Other Areas Figure 3.3a Percentage of respondents reporting a non-numeric ideal fertility preference by absolute poverty level, Central and Western Africa Figure 3.3b Percentage of respondents reporting a non-numeric ideal fertility preference by absolute poverty level, Eastern and Southern Africa Figure 3.3c Percentage of respondents reporting a non-numeric ideal fertility preference by absolute poverty level, South Asia, Southeast Asia, and Other Areas Figure 3.4a Modern contraceptive prevalence by absolute poverty level, Central and Western Africa Figure 3.4b Modern contraceptive prevalence by absolute poverty level, Eastern and Southern Africa Figure 3.4c Modern contraceptive prevalence by absolute poverty level, South Asia, Southeast Asia, and Other Areas Figure 3.5a Contraceptive method mix by absolute poverty level, Central and Western Africa Figure 3.5b Contraceptive method mix by absolute poverty level, Eastern and Southern Africa Figure 3.5c Contraceptive method mix by absolute poverty level, South Asia, Southeast Asia, and Other Areas Figure 3.6a Demand satisfied for modern methods by absolute poverty level, Central and Western Africa Figure 3.6b Demand satisfied for modern methods by absolute poverty level, Eastern and Southern Africa Figure 3.6c Demand satisfied for modern methods by absolute poverty level, South Asia, Southeast Asia, and Other Areas vii

10

11 PREFACE The Demographic and Health Surveys (DHS) Program is one of the principal sources of international data on fertility, family planning, maternal and child health, nutrition, mortality, environmental health, HIV/AIDS, malaria, and provision of health services. One of the objectives of The DHS Program is to provide policymakers and program managers in low- and middle-income countries with easily accessible data on levels and trends for a wide range of health and demographic indicators. DHS Comparative Reports provide such information, usually for a large number of countries in each report. These reports are largely descriptive, without multivariate methods, but when possible, they include confidence intervals and/or statistical tests. The topics in this series are selected by The DHS Program in consultation with the U.S. Agency for International Development. It is hoped that the DHS Comparative Reports will be useful to researchers, policymakers, and survey specialists, particularly those engaged in work in low- and middle-income countries. Sunita Kishor Director, The DHS Program ix

12

13 ABSTRACT Broad gains in contraceptive access and use have been made in low-income countries over the past decade while poverty has declined, but the trends have been uneven. In light of the Family Planning 2020 (FP2020) goals to improve modern contraceptive uptake, and the Sustainable Development Goals emphasis on equitable progress, there is renewed interest in monitoring fertility preferences and family planning outcomes by poverty level. However, studies of this topic are typically constrained by the fact that standard poverty measurements are relative within surveys and cannot be compared across countries or over time. This study develops and uses a measure of absolute poverty in 31 of the 69 FP2020 focus countries, employing both an unsatisfied basic needs approach and an asset index to help differentiate among the levels of the extremely poor. The measure of absolute poverty enables us to compare and test outcomes among comparable poverty groups both within and across countries. The study classifies married women into one of four absolute poverty groups based on their housing characteristics, household level of education, and assets. We compare results from the most recent Demographic and Health Survey in each of the 31 selected countries with results from an earlier survey in each country, conducted on average 10 years earlier. The study found a statistically significant and in many cases substantial decline in absolute poverty among married women in all 31 countries. There was wide variation in all key indicators across countries. On average, the ideal number of children declined most substantially among the poorest group of women, both in absolute terms and relatively across the decade. In the majority of countries there were statistically significant increases in modern contraceptive prevalence, demand satisfied for modern methods, and use of long-term versus short-term modern methods. Increases in all three indicators were greatest and most statistically significant among the poorest women. On average, inequalities between the non-poor and the poorest women declined, but substantial disparities by absolute poverty group remain both within and across countries. To address these disparities, we recommend further analysis incorporating background characteristics and programmatic case studies from countries that have largely achieved a high level of demand satisfied for modern methods while also increasing equity among poverty groups. KEYWORDS: poverty, modern contraceptive use, demand satisfied for modern methods, ideal family size, FP2020 xi

14

15 1 INTRODUCTION 1.1 Study Objectives Recognized as a highly cost-effective development intervention, family planning empowers women and couples to shape their own lives, supports healthier families, and helps to reduce poverty by increasing opportunities for economic growth (Alkema et al. 2013; Bongaarts et al. 2012; Carr et al. 2012; FP b; UNFPA 2017; United Nations, Department of Economic and Social Affairs, and Population Division 2017b). If all unmet need for modern contraception in developing countries were fulfilled, the number of unintended pregnancies, unplanned births, and induced abortions would decline by about 75%, and the resulting health benefits would be substantial, including far fewer maternal deaths (Guttmacher Institute 2017). In most developing countries, however, women in the bottom 20% of households by wealth, and particularly women in rural areas, are far less likely to have access to contraceptives than wealthier women and urban residents (UNFPA 2017). Too often, the poor are being left behind and losing out in access to quality health care and other essential services (United Nations 2018). Despite tremendous progress worldwide in boosting overall prosperity and in reducing extreme poverty, gaps in wealth have grown, and stark economic disparities remain. According to the most recent comprehensive data on global poverty, in 2013, an estimated 767 million people are living below the international poverty line of US$1.90 per person per day. By this standard, nearly 11% of the global population is poor, over half of whom are in sub-saharan Africa and another third in South Asia (World Bank 2016). The Sustainable Development Goals (SDGs) have placed emphasis on reductions in inequality, and in disaggregating outcomes by several categories including income (United Nations 2017). These goals, along with those of FP2020, have spurned interest in measuring and monitoring inequality in family planning outcomes. However, to date, research on fertility preferences, family planning, and poverty has been broadly constrained by the fact that most nationally representative surveys that produce these indicators, including the Demographic and Health Surveys (DHS), measure poverty in relative terms. The DHS Wealth Index (Rutstein and Johnson 2004) is widely used to compare relative economic status, based on household assets, construction materials, and services. The index was based on a methodology developed by Filmer and Pritchett (2001) to measure relative economic standing within a country at a given point in time in the absence of data on income. Using principal components analysis, households are scored relative to each other and the household population is divided into quintiles from richest to poorest. These scores and quintile rankings enable researchers to measure relative inequality in health outcomes at different points in time, but respondents cannot be compared in their economic status across countries or over time. The wealthiest 20% of the household population in a poor country may not be anywhere near what would be considered wealthy in their actual standard of living; conversely, in an affluent country the poorest may not be extremely poor by global living standards. Absolute measures of poverty and their relationship with family planning outcomes are the focus of this report. The study seeks to answer a few key questions. First, how do fertility intentions differ by levels of absolute poverty, and how has this changed over time? Second, are recent gains in modern contraceptive prevalence similar among the extremely poor, the poor, and the non-poor? Third, in what countries are 1

16 women in extreme poverty faring best and worst in terms of the percentage of demand for family planning satisfied by modern contraceptive methods? Having developed a measure of absolute poverty using DHS data for this report, we also examine levels and trends in poverty composition over time and across countries. 1.2 Background Sexual and reproductive health is an internationally agreed human right, endorsed by 179 governments in the 1994 Program of Action of the International Conference on Population and Development (Barot 2014; UNCESCR 2000). Since then, access to services for sexual and reproductive health has increased worldwide, and, more recently, global support for these rights has expanded significantly through the Family Planning 2020 initiative (FP2020) (United Nations, Department of Economic and Social Affairs, and Population Division 2015). FP2020, an outcome of the 2012 London Summit on Family Planning, is a global movement with an overall goal of reaching 120 million additional users of modern contraceptive methods in the world s poorest countries by 2020 (FP a). Substantial financial commitments to support this ambitious effort have been made by many FP2020 countries, donors, civil society organizations, and private-sector partners, and over 20 national governments have made commitments to address the policy, financing, delivery, and sociocultural barriers that prevent many women from accessing contraceptive information, services, and supplies. In 2017, it was estimated that in the 69 FP2020 focus countries more than 309 million women and girls were using a modern method of contraception; this is 38.8 million more than were using contraception in 2012, when FP2020 was launched an increase that is approximately 30% above the historical trend (FP b). Uptake of modern contraception has been improving in a majority of countries, across urban and rural areas, and at all levels of household wealth (UNFPA 2017). This growth in contraceptive use has led not only to improvements in health-related outcomes, such as reduced maternal mortality and infant mortality, but also to improvements in schooling and economic outcomes, especially for girls and women (United Nations, Department of Economic and Social Affairs, and Population Division 2017b) Fertility preferences Fertility preferences are an important precursor to the decision to use family planning; as desired family size declines, demand for contraception rises. Historically, over the course of the transition from high to low fertility, and as fertility entered the realm of conscious choice, desired family size has declined. During the demographic transition, economic and social changes such as industrialization, urbanization, and increased opportunities for education have led to lower levels of mortality and fertility (Bongaarts et al. 2012). The rising costs of raising children and their declining economic value for labor and old-age security are thought to have been central to the historic decline in desired family size, which in turn has led to growing demand for contraception (Becker 1960; Schultz 1973). Moreover, as countries develop, parents seek better health care, education, and opportunities for their children. Providing the means to do so increases the cost of raising many children, contributing to a desire for smaller but higher-quality families (Becker 1981). As might be expected, in countries with high levels of desired fertility, actual fertility also tends to be high, with few births averted by contraception; in contrast, where desired family size is low, many births are averted by contraception and actual fertility is lower (Bongaarts et al. 2012). While composite family 2

17 planning measures such as unmet need for family planning and the percentage of demand satisfied for contraception rely on women s immediate or retrospective fertility intentions, an historic relationship exists between modern contraceptive prevalence (mcpr) and the mean ideal number of children, as reported by women in surveys (FP b). Many social and cultural norms may affect the stated ideal number of children. For example, cultural pronatalism, widespread messaging about child limitation, or religious beliefs can influence perceptions of ideal numbers. Additionally, ex-post rationalization may play a role in the response to this question: while DHS surveys ask men and women if they could go back to the beginning of their reproductive lives, there may be a tendency to state an ideal number of children that is equal to or greater than one s current number of children. Moreover, summarizing one s ideal number of children with a single number can be a complex process; in many cultures, particularly South and Southeast Asia, numeric preference often depends on the number of sons and daughters. And for some women and men, childbearing is not under the calculus of conscious choice, while for others it is simply difficult to provide a numeric response when contingencies and complexities exist. DHS surveys allow for non-numeric responses; previous studies have shown that women who give non-numeric responses are likely to want more children than those who provide numeric responses, and are less likely to adopt behaviors that result in smaller families (Olaleye 1993; Riley, Hermalin, and Rosero-Bixby 1993; Upadhyay and Karasek 2012). Non-numeric responses have become increasingly rare in recent DHS surveys in sub-saharan Africa and elsewhere (Frye and Bachan 2017). We examine both ideal number of children among those who provided a numeric response to this question, and the percentage of non-numeric responses Family planning: levels and trends In 2017, 58% of married or in-union women of reproductive age were using a modern method of family planning worldwide. In sub-saharan Africa the level was much lower, at 32%, but higher in Asia, at 61% (United Nations, Department of Economic and Social Affairs, and Population Division 2017b). In 2017, the percentage of demand for family planning satisfied by modern contraceptive methods was 78% worldwide, but with wide regional variations (United Nations, Department of Economic and Social Affairs, and Population Division 2017b). Progress has been especially rapid in Africa, where the proportion of the demand for family planning satisfied with modern contraceptive methods increased from 41% in 2000 to 56% in 2017 (United Nations, Department of Economic and Social Affairs, and Population Division 2017a). Some countries with particularly rapid progress since 2000 in both stimulating and meeting demand for family planning, including Ethiopia, Malawi and Rwanda, serve as informative cases for the potential pace of change in other countries if investment and attention to family planning could be intensified. Most increases in contraceptive prevalence between 1990 and 2010 were attributable to rising use of modern methods, while the proportion of married or in-union women using traditional methods declined, from 6% in 2000 to 5% in 2017 (Alkema et al. 2013; United Nations, Department of Economic and Social Affairs, and Population Division 2017b). Method-specific contraceptive prevalence varies widely across the world. Overall, in Africa and Europe short-term and reversible methods, such as the pill, injectable, and male condom, are more common than other methods, whereas long-acting or permanent methods, such as sterilization, implants, and the IUD, are more common in Asia and Northern America (United Nations, Department of Economic and Social Affairs, and Population Division 2015). In 2015, over half of all contraceptive users worldwide relied on either 3

18 female sterilization (30%) or the IUD (21%), in large part due to patterns of long-term contraceptive use in China and India (United Nations, Department of Economic and Social Affairs, and Population Division 2015). Since 1994, the worldwide method mix has shifted away from female and male sterilization and toward injectables and male condoms. These shifts in part reflect the changing geographic composition of users over the past two decades, as contraceptive use has taken off in sub-saharan African countries where injectables are a common method (Bertrand et al. 2014). For the world as a whole, the share of total contraceptive use by the pill, implants, IUD, vaginal barrier methods, rhythm, and withdrawal has remained relatively stable over the past 20 years (United Nations, Department of Economic and Social Affairs, and Population Division 2015). The FP2020 initiative focuses solely on modern contraceptive methods, which have higher efficacy than traditional methods (Polis et al. 2016; Staveteig, Mallick, and Winter 2015). While family planning advocates have applauded the initiatives to increase access to modern contraceptive methods, they have also expressed concern that in the rush to meet the FP2020 goals, issues of voluntary use, reproductive choice, quality of care, and client-centered service delivery could be compromised (Bertrand et al. 2014; Hardee et al. 2013). In this paper we focus on levels of modern contraceptive use, method mix, and demand for family planning satisfied with modern methods. In addition to contraceptive prevalence, monitoring changes in method mix is important; providing access to a wide variety of modern methods makes it more likely that women can choose a contraceptive method that best suits their needs and preferences, thereby increasing consistency in use and minimizing discontinuation rates (Jain 2016; Jain et al. 2013). Contraceptive method mix highlights which methods are driving contraceptive use and can assist in identifying potentially underused methods (Bertrand et al. 2014). However, it is a complex indicator, as the choice of a contraceptive method reflects individual preferences, societal and cultural norms, and local and regional issues affecting contraceptive availability and accessibility (FP b). The dominance of a single method in a country may signal deficiencies in access to a full range of contraceptive methods (Bertrand et al. 2014; FP b; Ross, Keesbury, and Hardee 2015). Additionally, low rates of use among longer-term methods such as implants, injectable contraceptives, and intrauterine devices (IUDs) may be due to a shortage in human resources rather than the actual product itself. Of the 57 countries that have chronic shortages of human resources for health care, 36 are in sub-saharan Africa (WHO 2012). Expanding access to contraception requires increasing the supplies of quality contraceptive methods available and providing information about their safe use, as well as eliminating geographic, social, and economic barriers to contraceptive use (UNFPA 2017). In an attempt to offset the lack of trained health-care workers, several low- and middle-income countries have begun task-shifting, which can be defined as a more rational distribution of tasks and responsibilities among cadres of health workers, as community health workers (CHWs) have been recognized as an effective option for the delivery of more complex family planning services (Scott et al. 2015). The measure of demand for family planning met with modern contraceptive methods represents the percentage of women currently using a modern method among all women who are using or who have an unmet need for modern family planning (Bradley et al. 2012). This measure reflects voluntarism and informed choice it sets neither contraceptive prevalence nor fertility targets, but rather highlights the imperative to fulfill individuals and couples own choices with regard to number and timing of children (United Nations, Department of Economic and Social Affairs, and Population Division 2017b). It reflects 4

19 the existence of substantial levels of unmet need for family planning women who say they want to avoid childbearing but are not using contraception as well as women already using contraception to avoid pregnancy. Furthermore, the measure s focus on modern contraceptives reflects prioritization of these more effective methods. Modern contraceptive use, which results in fewer unintended pregnancies compared with traditional methods, can help individuals and couples achieve their reproductive intentions (Fabic et al. 2014). Across low-income countries, women in the top wealth quintile have 50% more demand for family planning met with modern contraceptive methods, on average, compared with all wealth quintiles combined (Fabic et al. 2014). Nevertheless, the historical experiences of formerly low-income countries, such as South Korea and Thailand, indicate that with focused attention and widespread support, the demand for family planning met with modern contraceptives can increase from low levels to as high as 75% in 20 years or fewer (Fabic et al. 2014; Robinson and Ross 2007). In the same group of countries, the percentage of demand for family planning satisfied with modern contraceptive methods increased from 59% in 2000 to 68% in 2017 (United Nations, Department of Economic and Social Affairs, and Population Division 2017a) Global poverty: levels and trends In 2013, an estimated 11% of the world s population, about 767 million people, were living under the international poverty line of US$1.90 a day, down from an estimated 12% in 2012 (World Bank 2016). When measured in all of its dimensions, progress in poverty reduction and shared prosperity over the past three decades has been significant (Cruz et al. 2015). Since 1990, nearly 1.1 billion people have moved out of extreme poverty (UNDP 2016). Much of this reduction has been driven by remarkable progress in the East Asia and Pacific region with 71 million fewer poor people, notably in China and Indonesia, and in South Asia with 37 million fewer poor, notably in India (World Bank 2018). A significant change in the geography of poverty has meant that in 2013 sub-saharan Africa contained more than half the world s poor. This is despite the fact that the African subcontinent experienced progress in lowering both the percentage of the population that are poor (by 1.6 percentage points) and the number of poor (by 4 million in ) (World Bank 2016). These achievements are modest, however, compared with East Asia and Pacific and South Asia. Other regions with lower poverty rates and total numbers notably Eastern Europe and Central Asia, as well as Latin America and the Caribbean saw marginal declines in poverty in (World Bank 2016). The Multidimensional Poverty Index (MPI) measures non-income dimensions of poverty, and aims to capture severe deprivations that are faced by individuals with respect to education, health, and living standards (UNDP 2016; United Nations 2016). While MPI has declined significantly, it remains unacceptably high in some areas. The continued persistence of geographically concentrated pockets of deep multidimensional poverty within many countries has led to conflicting views about the extent and pace of progress in poverty reduction (Cruz et al. 2015). Poverty is a major cause of ill health and is a barrier to accessing health services for many people. Inequality in access to health services is widespread in some countries and is associated with higher income inequality. Particularly, maternal health and adolescent fertility are closely related to income inequality and the incidence of poverty (Gonzales et al. 2015). Globally, the poor are predominantly rural, young, poorly educated, mostly employed in the agricultural sector, and live in larger households with more children 5

20 compared with the non-poor (Castañeda et al. 2016). Developing countries tend to exhibit wider withincountry inequality relative to developed countries (World Bank 2016) Family planning outcomes and poverty Many developing countries have improved their capacity to provide modern contraception and to reduce wealth-based inequality in satisfying the demand for family planning. In less-developed countries, however, there tends to be a wide gap in contraceptive use between households in the highest and lowest wealth quintiles (52% versus 35%) (UNFPA 2017). This gap has persisted despite general improvements in socioeconomic status and the expansion of family planning services worldwide (Creanga et al. 2011). In the majority of developing countries, contraceptive prevalence is lower among women who are poorer, rural, and less educated compared with richer, urban, and more educated women (UNFPA 2017, 2013). Low-income countries themselves vary substantially in modern contraceptive prevalence. In 2017 the range was from below 10% in Chad, Guinea, and South Sudan to 67% in Zimbabwe and 71% in Democratic People s Republic of Korea (United Nations, Department of Economic and Social Affairs, and Population Division 2017b). Poverty, of course, is not the only factor in access to family planning. Local and national service and policy environments, levels of education, age structure, and programmatic initiatives can increase family planning use even among the very poor. Fertility has declined rapidly in a few countries with unfavorable development conditions (e.g., Bangladesh, Indonesia, Nepal, and Sri Lanka). These are traditional, poor, rural, and agricultural societies, yet fertility has declined to low levels. The main explanation for these unexpected trends is the priority their governments have given to social development (e.g., schooling and women s empowerment) and the implementation of effective family planning and health programs. No fertility decline has been observed in a poor and largely illiterate country in the absence of a strong family planning program (Bongaarts et al. 2012). In 2017, in 76 out of 185 countries for which data are available, 75% or more of the total demand for family planning was met with modern contraceptive methods. These countries include 14 in Africa, 13 in Asia, 25 in Latin America and the Caribbean, and 24 in other regions. In contrast, in another 45 countries less than half of the total demand for family planning was met with modern methods. Among these countries, 32 were in Africa, 8 in Asia, 5 in Europe, and 3 in Oceania (United Nations, Department of Economic and Social Affairs, and Population Division 2017b). Among the FP2020 focus countries, the percentage of demand for family planning satisfied with modern methods is lowest in the four sub-regions of Africa, at 24% in Central Africa, 37% in Western Africa, 62% in Eastern and Southern Africa, and 66% in Middle East and Northern Africa, followed by Latin America and the Caribbean (66%), South Asia (72%), Southeast Asia and Oceania (75%), and Eastern and Central Asia (78%) (FP b). An analysis of data on the proportion of demand for family planning satisfied with modern contraception among women who are married or in a union shows that women in the least developed countries have less access than women in other developing countries. It also shows that regardless of a country s income grouping, the richest 20% of households on average have the most access, and the poorest 20% have the least access. There are exceptions, however, where use of family planning is generally more equitable. In Bangladesh, Bhutan, Cambodia, and Thailand, contraceptive prevalence is higher among the poorest 20% of households than the richest 20% (UNFPA 2017). In these and several other countries, concerted efforts to expand family planning coverage have led to almost universal access to modern contraception, and near- 6

21 equitable prevalence of contraceptive use across the wealth spectrum (UNFPA 2017). A recent study that standardized poverty measures in DHS surveys to assess the contributions of family planning programming versus changing living standards found that in a majority of study countries, expanding family planning services contributed more to an increase in contraceptive use than improvements in living conditions, across all deciles of wealth (Emmart, Winfrey, and Davis 2017). While some countries have made exceptional progress in reducing inequality of access to contraception, others have made great progress in expanding coverage of contraceptive services. Lesotho, Rwanda, and Sierra Leone are examples of countries that have made exceptional progress in both areas over about a 10- year period. For example, analysis by relative household wealth quintiles in Rwanda shows that a previous wide gap in demand for family planning satisfied by modern methods has been effectively closed at an access proportion of close to 70% among all five wealth quintiles in the later survey; Lesotho experienced similarly strong progress (UNFPA 2017). 7

22

23 2 DATA AND METHODS 2.1 Country and Survey Selection This study employed data from countries with nationally representative Demographic and Health Surveys (DHS) based on the following criteria: (1) The country was among the current list of FP2020 focus countries (FP ); (2) A standard DHS survey was conducted in 2012 or later that was available by May 1, 2018; (3) A DHS survey was available that was at least five years older than the more recent survey and was conducted after 1995; and (4) Both surveys included all variables necessary for the analysis of absolute poverty. If more than one older survey met the criteria for year and variables, we gave preference to the survey that was closest to a 10-year difference from the most recent survey. If two older surveys were equally close to 10 years, we gave preference to the earlier of the two. This strategy resulted in selection of 62 surveys in 31 countries, shown in Figure 2.1. Twenty-eight study countries were classified into one of three major world regions: Central and Western Africa, Eastern and Southern Africa, or South and Southeast Asia. Three additional countries that qualified for inclusion Egypt, Haiti, and Kyrgyz Republic do not share a common region with any other study countries and are henceforth referred to as belonging to an Other Areas category. Figure 2.1 Study countries 9

24 Our analysis focused solely on currently married women of reproductive age (15-49). Per standard DHS definitions, the term currently married means that the woman is married or living with a man as if married. The 62 surveys we study are shown in Table 2.1 along with the corresponding weighted sample sizes of married women. Surveys included in the study were fielded as early as 1996 and as late as 2016; intracountry gaps ranged from 5 years (Sierra Leone) to 16 years (Comoros). On average there was a 10-year difference between survey rounds. Table 2.1 Surveys included in the analysis Survey 1 Survey 2 Year Sample size Year Sample size Number of years between survey rounds 1 Central and Western Africa Chad , , DR Congo , , Ghana , , Guinea , , Liberia , , Mali , , Niger , , Nigeria , , Senegal , , Sierra Leone , , Eastern and Southern Africa Burundi , , Comoros , , Ethiopia , , Kenya , , Lesotho , , Malawi , , Rwanda , , Tanzania , , Uganda , , Zambia , , Zimbabwe , , South and Southeast Asia Cambodia , , India , , Indonesia , , Nepal , , Pakistan , , Philippines , , Timor-Leste , , Other Areas Egypt , , Haiti , , Kyrgyz Republic , , If survey fieldwork spans two years, it is assumed to have been fielded at the midpoint between those years, e.g., for a survey. 10

25 2.2 Key Outcome Indicators We assess five key outcome indicators in relationship to absolute poverty, as follows: Mean ideal number of children and non-numeric fertility preferences Toward the end of the DHS interview, women who have living children are asked, If you could go back to the time you did not have any children and could choose exactly the number of children to have in your whole life, how many would that be? Women who do not have living children are asked, If you could choose exactly the number of children to have in your whole life, how many would that be? Possible response options are none, a specific number, or an other response. If a non-numeric response is given, interviewers are instructed to probe for a numeric response. The mean ideal number of children refers to the mean of the number of children women specified in response to this question, among those who gave a numeric response. As the term implies, a non-numeric response means that, in spite of the numeric probe, the woman gave a qualitative response to this question, such as it depends, up to god, as many as possible, uncertain, and so forth. Typically, these responses are interpreted as fatalistic, indicative that fertility is outside the realm of conscious choice, although they may also suggest substantial uncertainty in some circumstances (Hayford and Agadjanian 2011). Contraceptive method mix Method mix refers to the composition of family planning method types among sampled women who report currently using a means to delay or avoid pregnancy. If women report more than one method, the method that is most effective is considered their primary method. For our purposes, we classified reported method types into four major groups: long-term modern methods, short-term modern methods, traditional methods, and folkloric methods, as follows: 1. Long-term modern methods: IUDs, implants/norplant; sterilization (male or female). 2. Short-term modern methods: Pill, injectables, condom (male or female), emergency contraception, basal body temperature, Billings ovulation, the sympto-thermal method, standard days methods, and other modern methods 1 such as diaphragms, cervical caps, foam, jelly, and suppositories. 3. Traditional methods: Periodic abstinence (the rhythm method), withdrawal (coitus interruptus), prolonged breastfeeding, douche. 4. Folkloric: Herbs, massage, other folkloric methods, and any other method not named above and not specifically classified as modern. 1 Other modern methods is a category available for interviewers to select in most DHS surveys. Prior to the survey, interviewers are instructed on which methods may be included in the category; however, the full list may not be available to analysts. 11

26 Modern contraceptive prevalence Modern contraceptive prevalence (mcpr) among married women is defined as the percentage of women who report that they or their sexual partners are currently using a modern method of family planning. Eligible methods include all long-term and short-term modern methods described above. Demand for family planning satisfied by modern methods Demand satisfied by modern methods (DSMM) is defined as the number of women who are currently using, or whose sexual partner is currently using, at least one modern contraceptive method as a proportion of the number of women of reproductive age who use any method of family planning or who have an unmet need for family planning (FP b; United Nations, Department of Economic and Social Affairs, and Population Division 2017b). 2.3 Absolute Poverty Measurement Rationale In the absence of income and expenditure data, researchers at the DHS Program have developed an economic status measure, the DHS Wealth Index (Rutstein and Johnson 2004), based on an earlier index developed using DHS data (Filmer and Pritchett 2001). The DHS Wealth Index is computed based on assets, services, and household characteristics within each survey. Other variables such as the presence of a domestic servant are added if the survey allows. This composite wealth index has proven enormously useful to the study of inequalities in health behaviors and the effect of relative economic status on health outcomes. Although commonly misunderstood as a proxy for income, which tends to be volatile in many low-income settings, the DHS Wealth Index is designed to capture a more stable measure of economic status than income alone, akin to Milton Friedman s concept of permanent income (1957). Despite the enduring value of the DHS Wealth Index, its key limitation is that measurement is relative for any given country at a point in time, based on the specific assets, services, and construction materials asked about in that survey and their distribution within the population. The principal components analysis used in computations assigns scores to assets based in part on their prevalence; as asset ownership becomes more widespread and as construction materials and access to household services such as electricity and running water improve, the scores assigned to these assets and services by the principal components index shift. For example, having a cellphone in an early survey might be an important indicator of wealth, but in a later survey, if cellphones have become nearly ubiquitous, the wealth score gained by owning a cellphone might be near zero. Hence, a household with a stable bundle of assets, services, and construction materials might be scored as wealthy in one survey and poor in another. Thus, while the DHS Wealth Index is enormously useful within countries, it is constrained by its specificity to a given country and time period Previous approaches Efforts to standardize the DHS Wealth Index across countries and over time are made challenging by the fact that earlier surveys asked relatively few questions that could be used to measure economic status. Before the late 1990s, DHS surveys typically only asked about assets directly related to a key health outcome, for example ownership of radios and televisions in relation to family planning messaging and use, or dirt floors and inadequate toilets in relationship to diarrhea among young children. As the DHS Wealth 12

27 Index progressed, the number of questions about household assets and services grew, and then in turn expanded further as countries became aware of the wealth index and wanted to capture information about salient assets. These developments have allowed the index to better differentiate households by economic status, particularly at the upper ends of the spectrum, but they make retrospective comparisons difficult. Researchers at The DHS Program have worked to standardize the wealth index, both across countries and time periods (Rutstein and Staveteig 2014) via the Comparative Wealth Index, and within countries over time via the Harmonized Wealth Index (Staveteig and Mallick 2014). The Comparative Wealth Index, which uses an anchoring approach, enables greater cross-survey comparability but occasionally suffers from undesirable distortions induced by linear displacement (Staveteig and Mallick 2014). The Harmonized Wealth Index produces a truly comparable wealth score between surveys that is unaffected by displacements. By making use of assets that are salient and sometimes specific to a given country, such as a wardrobe in Bangladesh or a water heater in Egypt, the Harmonized Wealth Index is more accurate than a cross-country common-denominator approach; but its corollary drawback is that such specific assets and within-country distributions inhibit cross-country comparisons. Ultimately, any common-denominator approach to remaking a standard wealth index has difficulty differentiating between households at the upper ends of the wealth scale, due to the limited number of asset questions in early surveys. Instead of trying to create a comparable wealth index, this paper, with its focus on poverty, instead creates a standardized poverty measure aimed at differentiating from among the lowest levels of the wealth index; it groups the non-poor into a single reference group Our approach Inspired by Amartya Sen s seminal work on measuring poverty in terms of absolute, not relative deprivations (1976, 1982), we developed for this paper a direct method of poverty measurement: we measure a household s achievement of basic needs to assess what standard of living a household actually affords. This approach is in some ways preferable to using monetary income or wealth as an intermediary variable, as market prices for basic necessities can vary widely by country. Consider, for example, the amount of monetary income required for a household located in an area with accessible electrical lines and plentiful piped drinking water to access those services, versus a second household in a rural area of a developing country with virtually no public infrastructure beyond roads and schools. In the latter scenario clean water and electricity could be obtained, perhaps through generators and bottled water shipments, but only at a very high price. As such, simple income or monetary wealth comparisons between these two households would be insufficient to gauge deprivations of basic needs. Our approach follows a line of earlier work on multidimensional poverty measurement using an index of unsatisfied basic needs (UBN). This framework, often referred to in the literature by its Spanish name Indice de Necesidades Básicas Insatisfechas, was formalized by the U.N. Economic Commission for Latin America (ECLAC) and the Census Institute in Argentina in the 1980s (Feres and Mancero 2001; Instituto Nacional de Estadística y Censos [INDEC] 1984). The UBN was designed to capture dimensions of poverty that could be determined from census data and that would be difficult to observe from income alone. It originally aimed to measure human deprivations, but over time other nonmonetary aspects of poverty such as household crowding and children s non-enrollment in school, which were associated with poverty were added to the measure. The index is now widely used across Latin America (Feres and Mancero 2001). Although there is no single definition of unsatisfied basic needs, the index typically 13